In the regular swell of
political dialogue that comes with the continuous election cycle, I find myself
considering what it means to vote in our massive political system. Intended to be
responsive to the popular changes in political sentiment over time, given so
many voters, so many opinions with equal weight as mine, it is hard to perceive
my ballot as anything more than negligible. In a group of this scale, my vote
is not the tipping point. My voice is easily crowded out and smothered.
As a student of biology
trained to sift out motif, when describing a behavioral or ecological
phenomenon I often indulge in analogy. In nature, many animals have “voting
power” in their social groups. Generally speaking, information from multiple
individuals is summed somehow to produce decisions for the whole. Choosing
where to nest, choosing where to forage, coordinating a hunt, migrating to
distant places –these group behaviors are executed in nature through a surprising
number of different information distribution strategies. When choosing between
two political candidates, our system assigns power uniformly across individuals
within a group, but other animals don’t always make group decisions like this.
Sometimes only a few voters are needed, sometimes some voters count more than
others, sometimes whole-group communication is possible, and sometimes
group-level decisions must be made with only local information.
In the Bee Lab we tend to
focus on social insects; but social mammals, flocking birds, and schooling fish
must also coordinate with their neighbors and respond to their environment to
operate as a whole. To help understand the similarities and differences in how
different species reach consensus to coordinate behavior, I found a review
by Conradt and Roper to be useful. In order to place the continuum of group
behavioral strategies into a more elegant framework, the article identifies
factors like communication ability, group size, and conflict of interest within
a group as particularly important in predicting how voting power is assigned
and exercised for a particular species.
In terms of distribution of
voting power, perhaps the most egalitarian, unbiased mechanism for forming
consensus is expressed in a colony of social insects. Where human equality is
ensured only by social policy (fraught with irregular application and variable
enforcement), insect equality is firmly grounded in physiology.
Bee and ant societies are
such that no individual is ever aware of the state of the colony as a whole,
and individual behavior is dictated purely by the chemical and tactile feedback
from its immediate surroundings and neighbors. Social insects don’t really
operate “democracies” like our own because the individual votes are never
completely pooled or tallied. The entire group must remain cohesive for certain
critical tasks like nest relocation–a complicated feat of synchrony for such a
large number of small brains.
 |
A colony in a hollow tree, a commonly chosen nest site for wild honeybees [1]
|
An article
by Seeley, Vissher and Passino describes the honeybee mechanism of what we may
call “voting” in reaching consensus for nest relocation: individual scouts find
potential nest sites and then communicate their location using the waggle
dance.
How does the group decide on
a single nest site out of multiple options, and how well do they choose their
new site? The mechanism observed by Seeley is not only beautiful, but also
effective. Most notably, it maintains global functionality with only local
feedback.
Nest quality across multiple
sites is assessed using a feedback loop, manifested in the rates of local
interactions. Scouts fly from the hive in search of suitable nest sites,
directly inspecting hollow tree trunks or rock crevices. Scouts that find a
REALLY GOOD site dance more, increasing
the rate of workers searching for their particular site. If different locations
are deemed suitable by different scouts, the colony will allocate workers to
both sites simultaneously. If enough scouts aggregate at one site—reaching a
quorum—the bees will fly back to the hive to catalyze a move by impeding other dancers, speeding up the “decision.” To me, the genius of the
process and perhaps a key achievement in collective intelligence is how
honeybees overcome indecision and effectively
choose between multiple viable options without any individual ever
experiencing more than one.
A social insect colony is responsive
to environmental fluctuation not through the relatively rigid rules of
individual interactions, but rather by the fluidity in the changing rates of
said interactions. To witness a honeybee colony with the locations of multiple
sites being waggled simultaneously is to witness a colony weigh its options and
make up its mind.
According to Conradt’s
framework, groups of ants and honeybees have effectively zero conflict of
interest amongst individuals. “Voting” in insect societies is distributed; they
lack hierarchy in the sense that a stimulus from a single worker is equal in
weight to a stimulus from any other worker. A colony contains only one
reproductively active member, and is effectively one unit from the evolutionary perspective. Sterile sisters can afford to invariably trust one another. Economically, the
colony avoids the costs associated with conflicts of interest to allocate colony
members to go where they’re most needed, quickly.
In operations like nest
relocation, fascinating self-organizing principles are at work that allow a
colony to respond to a chaotic environment. To preplan and coordinate full
scale evacuation without any central control and with relatively rigid rules of
interaction between individuals is remarkable. For a bee colony, a group stays
aware and responsive through complex feedback loops that allow the collective
biomass to allocate itself more efficiently. To me, that’s cool.
But looking at a single
worker it’s hard to imagine the potential inherent in each antennal contact,
the information being conveyed when a worker palpates the air for pheromone. The
brevity and sheer number of these micro contacts masks their significance. That
the entire behavioral spectrum of a colony, its full potential to respond to
environmental change is regulated and determined at the tap of an antenna, the
waggle of an abdomen.
 |
| An individual worker, the agent of colony function [2] |
Additional Reading:
Conradt, L and Roper, T.J.
(2005) Consensus decision making in animals. Trends in Ecology and Evolution 20, 449-456
Seeley, T.D. (1989) The
Honeybee Colony as a Superorganism. American
Scientist 77, 546-553
Seeley, T.D. and Visscher,
P.K. (2004) Group decision-making in nest site location by honeybees. Apidologie 35, 101-116
Seeley, T.D. et al. (2011) Stop Signals Provide Cross Inhibition in Collective Decision-Making
by Honeybee Swarms. Science 335,
108-111
Strandberg-Peshkin, A. et al. (2015) Shared decision-making
drives collective movement in wild baboons. Science
348, 1358-1361
Media Credits:
[1] Wild Honeybee Colony:
Photo by Alex Wild
[2] Individual Bee:
Photo by Alex Wild
Really cool! Plato and Aristotle compared political organization to beehives (for better or for worse), and I think your analogy brings up interesting questions about individual stakes and the best chance for survival of the greater system. Each individual bee only stands to lose or gain as much as its colony so there are no competing interests, and the waggle communication (and stop signal system you mention?) is a consistent and reliable indicator of the whole colony’s potential gains/losses. Sounds like the opposite of politics to me, where a politician or voter’s interests are often decoupled from the “system”’s interests (if there actually is one overall “system” interest at all) – even if one voter=one vote (although lots of campaign money might be worth more than a vote...). And ironically it’s Plato’s beehive-inspired-sociopolitical-castes theories that reinforce the existence of competing interests in one society. Maybe people are not bees?
ReplyDeleteOne great place to read more about the similarities & differences between human and honeybee voting is Tom Seeley's fascinating book, "Honeybee Democracy".
ReplyDeleteI wouldn't have thought of this to be a linear voting process ... but maybe the quorum effects make up for it!
ReplyDeleteChaitanya, can you explain what you mean? Are you talking about the time complexity of the voting algorithm? It is definitely the quorum that determines how fast it goes (Check out the Seeley 2011 paper for a nice model). As you can imagine, you can make it faster by making the quorum smaller, but then mistakes are more likely. This is demonstrated in ants in Franks et al. 2003 "Speed versus accuracy in collective decision making".
Delete